Developing and Utilizing a Large-Scale Cantonese Dataset for Multi-Tasking in Large Language Models